The transcription factor FLC confers a flowering response to vernalization by repressing meristem competence and systemic signaling in Arabidopsis

Abstract

Floral development at the Arabidopsis shoot apical meristem occurs in response to environmental cues that are perceived in different tissues. Photoperiod is detected in the vascular tissue of the leaf (phloem) and promotes production of a systemic signal that induces flowering at the meristem. Vernalization, the response to winter temperatures, overcomes a block on photoperiodic floral induction. In Arabidopsis, this block is caused by inhibitors of flowering that comprise several related MADS-box transcription factors, the most prominent of which is FLC. We show that FLC delays flowering by repressing production in the leaf of at least two systemic signals, one of which is controlled by the RAF kinase inhibitor-like protein FT. Reducing expression of these signals indirectly represses expression of genes involved in floral induction at the meristem. In addition, FLC expression in the meristem impairs response to the FT signal by directly repressing expression of the SOC1 MADS-box transcription factor and preventing up-regulation of the bZIP transcription factor FD. Repression of genes acting at multiple levels in this hierarchy is required for the extreme delay in flowering caused by FLC. An FLC:HA fusion protein binds directly in vivo to the promoter regions of FD and SOC1 and to the first intron of FT. Thus vernalization relieves transcriptional repression of key regulatory genes in both the leaf and meristem, allowing production of systemic signals in the leaves and conferring competence on the meristem to respond to these signals.

Figure 1.

Effect on flowering time of misexpression of FLC from heterologous promoters. (A) Flowering times expressed as total leaf numbers of flc-3 transgenic plants in which FLC is expressed from tissue-specific promoters. All plants were grown under LDs (16 h light/8 h dark). (B) As A, but plants were grown under SDs (10 h light/14 h dark). At least 10 transformants were recovered for each construct, and the flowering time of a representative transformant for each construct is shown.

Figure 2.

Analysis of mRNA levels in leaves of different genotypes by RT–PCR. (A) FT mRNA abundance in RNA extracted from leaves of FRI flc-3, KNAT1∷FLC flc-3, SUC2∷FLC flc-3, and FRI FLC plants grown under LDs. (B) SOC1 mRNA abundance in RNA extracted from leaves of FRI flc-3, KNAT1∷FLC flc-3, SUC2∷FLC flc-3, and FRI FLC plants grown under LDs. Leaves were harvested at dusk.

Figure 3.

Effects of misexpression of SOC1, FT, and FLC in different genetic backgrounds on flowering time. (A) Flowering times are expressed as total leaf numbers of soc1-1 transgenic plants in which SOC1 is expressed from tissue-specific promoters. All plants were grown under LDs. (B) As A, but plants were grown under SDs. (C) Flowering times of transgenic plants misexpressing FLC and either FT or SOC1. Plants were grown under LDs. Some genotypes involve crosses of Columbia FRI flc-3 to Ler, but because the Ler FLC allele does not respond to FRI, no activation of endogenous FLC occurs (Michaels and Amasino 1999). (D) Flowering times of transgenic lines expressing SOC1 from tissue-specific promoters in co-2 or soc1-1 and carrying the ft-7 mutation, as well as double mutants carrying ft-7 and other flowering-time mutations. Plants were grown under LDs. (E) Analysis of FT mRNA abundance by RT–PCR in the leaves of Ler, co-2, SUC2∷SOC1 co-2, and SUC2∷CO co-2 grown under SDs. Leaves were harvested at dawn.

Figure 4.

Analysis of SOC1 expression in the meristem by in situ hybridization. (A) Ler plants were grown under 14 SDs (8 h light) and shifted to LDs (16 h light) at dawn. This plant was harvested 8 h after dawn, which represented the end of the SD treatment or Time 0 (left). (B) This plant was harvested from the same experiment as A but at dawn the next day, 16 h after the end of the last SD. SOC1 expression was detected in the meristem. (C,D) Columbia (Col) wild type grown under LDs for 6 d (C) and 10 d (D). SOC1 expression is detectable in the 10-d-old plant. (E,F) Meristem of SUC2∷FLC flc-3 (Col) plants after 6 LDs (E) and 10 LDs (F). SOC1 expression is not detected. (G,H) ft-7 (Ler) plants after 10 LDs (G) and 20 LDs (H). SOC1 expression is only detected weakly after 20 LDs. (I,J) Meristems of Ler (I) and SUC2∷CO co-2 (J) after 10 LDs. (K,L) SUC2∷CO ft-7 co-2 after 10 LDs (K) and 16 LDs (L). All SUC2∷CO co-2 transgenic lines are in a Ler background. Bar, 100 μM.

Figure 5.

Analysis of FD and FLC expression in the meristem by in situ hybridization and RT–PCR. (A) FD expression in the meristem of Col wild-type, KNAT1∷FLC flc-3, and SUC2∷FLC flc-3 plants after 6 LDs (left column), 8 LDs (middle column), and 10 LDs (right column). FLC transgenic lines are in a Col background. Bar, 100 μM. (B) Analysis of FD mRNA level by RT–PCR in FRI flc-3, KNAT1∷FLC flc-3, FRI FLC plants after 10 LDs. (C) FD expression in the meristem of ft mutants after 6 LDs (left), 8 LDs (middle), and 10 LDs (right). (D) Before vernalization, FRI FLC plants were grown under 14 SDs (top left), then vernalized for 28 d, and immediately after vernalization (top right), FLC expression was analyzed in the meristem by in situ hybridization. FD expression before vernalization (bottom left) and after vernalization (bottom right) of the same plants.

Figure 6.

Analysis of SOC1 expression in the meristem by in situ hybridization. (A) SOC1 expression in the meristem of wild-type Col and KNAT1∷FLC flc-3 plants grown for 6 and 10 LDs. SOC1 expression is detectable in wild-type (WT) but not KNAT1∷FLC after 10 d. (B) SOC1 expression in the meristem of Ler and fd-1 (Ler) after 8, 10, and 12 LDs. (C) Columbia and fd-3 (Col) mutant plants were grown under SDs (for 14 d) and shifted to LDs at dawn. Eight hours after dawn was the end of the SD treatment and was designated as the zero time point (0, left). Sixteen hours later, SOC1 expression was detected in the meristem of Col but not the fd-3 mutant (16, middle). Twenty-four hours latter, SOC1 expression was still not detected in the fd-3 mutant (24, right). Bar, 100 μM.

Figure 7.

ChIP analysis of SOC1, FT, and FD loci. (A) A SOC1 promoter fragment containing the FLC-responsive CArG box was amplified from DNA immunoprecipitated with control (anti-Rat IgG) or anti-HA antibodies from Ler or 35S∷FLC:HA Ler. A twofold dilution series of the input DNA was amplified as a semiquantitative standard. The SOC1 promoter region assayed is indicated by a thin horizontal line. Coding regions are indicated by shaded boxes. (B) Five FT fragments spanning introns 1 and 2 were amplified from immunoprecipitated DNA as described in A. Fragment 3 contains a putative FLC-responsive CArG box and was amplified from immunoprecipitated DNA, and twofold dilutions of the input DNA are shown as a control. Horizontal lines numbered 1–5 indicate assayed fragments. (C) Quantification of the amplified signal from fragments 1–5 relative to the input. (D) Four fragments spanning the FD promoter region were amplified from immunoprecipitated DNA as described in A.

Figure 8.

A schematic diagram illustrating the interactions between FLC and the photoperiod pathway in the leaf and meristem and the effect of vernalization on these interactions. Prior to vernalization (diagram on left), FLC acts in the leaf to repress transcription of FT and of other genes that are illustrated as X. In the meristem, FLC represses transcription of FD and of SOC1. Vernalization reduces FLC expression both in the leaf and the meristem. Removal of FLC from the leaves causes FT and X to be expressed. FT controls a systemic signal (shown in blue) that is required for SOC1 expression at the meristem. FT mRNA was recently proposed to move from the leaves to the meristem during floral induction (Huang et al. 2005), and therefore this signal might be comprised of FT mRNA. FD is also required at the meristem for SOC1 expression, and because FT and FD interact (Abe et al. 2005; Wigge et al. 2005), this heterodimer might activate SOC1 expression directly or indirectly. FD expression at the meristem also rises during floral induction. The increase in FD expression does not require FT but is blocked by expression of FLC in the leaves. Therefore we propose that in the leaves, FLC blocks expression of X, which controls a second systemic signal (shown in red) that increases FD expression at the meristem. The reduction of FLC expression in the meristem during vernalization allows FD and SOC1 expression in the meristem to rise in response to the X signal and FT/FD, respectively.